A Reliable Spectrum Sensing Method in the Presence of Malicious Sensors in Distributed Cognitive Radio Network

Spectrum sensing is an important issue in cognitive radio networks. But due to noise and shadowing effects, the sensing results of sensors may be erroneous. This will degrade the performance of the decision making process. To solve this problem, some sensors may cooperate with each other to decide about the channel status. Cooperative sensing will improve the performance of the system. But, some of the sensors may be malicious sensors. These sensors behave egotistic in the system by sending incorrect reports. Therefore, cooperative sensing is susceptible to some attacks, which may disrupt the sensing process. In this paper, we propose a reliable method, which is based on clustering the cooperating sensors. A cluster with no malicious sensor will be found by a fast searching algorithm. Unlike most of the previous works, which focus on detecting the malicious sensors, our model is based on detecting the trusted sensors. The proposed model detects a cluster with no malicious sensor and uses the sensing results of this cluster. Therefore, the sensing information of the malicious sensors will not being participated in the spectrum decision process. In addition, this method has high reliability in detecting the channels; it is robust to noise and shadowing effects. Moreover, the raw data that are exchanged between CRs and fusion center reduces significantly in our model compared with other approaches.